• Laser & Optoelectronics Progress
  • Vol. 56, Issue 7, 071004 (2019)
Shuyu Huang and Qingbing Sang*
Author Affiliations
  • Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computational Intelligence, College of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China
  • show less
    DOI: 10.3788/LOP56.071004 Cite this Article Set citation alerts
    Shuyu Huang, Qingbing Sang. No-Reference Stereo Image Quality Assessment Based on Image Fusion[J]. Laser & Optoelectronics Progress, 2019, 56(7): 071004 Copy Citation Text show less
    Framework of stereo image quality assessment algorithm based on fused image and convolutional neural network
    Fig. 1. Framework of stereo image quality assessment algorithm based on fused image and convolutional neural network
    Left and right views and fused image of stereo image. (a) Left view; (b) right view; (c) fused image
    Fig. 2. Left and right views and fused image of stereo image. (a) Left view; (b) right view; (c) fused image
    Results with normalized luminance coefficient. (a) Fused image with normalized luminance coefficient;(b) probability density distribution
    Fig. 3. Results with normalized luminance coefficient. (a) Fused image with normalized luminance coefficient;(b) probability density distribution
    Architecture of convolutional neural network
    Fig. 4. Architecture of convolutional neural network
    Predicted scatter plots. (a) LIVE 3D Phase I database; (b) LIVE 3D Phase II database
    Fig. 5. Predicted scatter plots. (a) LIVE 3D Phase I database; (b) LIVE 3D Phase II database
    MethodRSROCCRPLCC
    LP0.94620.9687
    NSCT0.94550.9688
    GFF0.94830.9642
    DWT0.94290.9675
    Table 1. RSROCC and RPLCC of different fused methods on LIVE 3D Phase I database
    MethodRSROCCRPLCC
    LP0.92240.9237
    NSCT0.92180.9243
    GFF0.89550.9072
    DWT0.93420.9428
    Table 2. RSROCC and RPLCC of different fused methods on LIVE 3D Phase II database
    MethodRSROCCRPLCC
    LP0.79400.7763
    NSCT0.78280.7765
    GFF0.80120.7921
    DWT0.80330.7822
    Table 3. RSROCC and RPLCC of different fused methods on MCL-3D database
    MethodJP2KJPEGWNBLURFFAll
    PSNR0.7990.1210.9320.9020.5870.834
    SSIM0.8580.4360.9380.8790.5860.876
    Method in Ref. [1]0.9100.6030.9290.9310.6990.899
    Method in Ref. [10]0.8750.6140.9430.9370.7810.921
    Method in Ref. [9]0.8660.6750.9140.5550.6400.383
    Method in Ref. [8]0.8630.6170.9190.8770.6520.891
    Method in Ref. [15]0.8890.6130.9090.8770.7580.925
    Method in Ref. [14]0.8370.6380.9310.8330.6490.892
    Method in Ref. [16]0.9170.7820.9100.8650.6660.911
    Proposed0.7860.9290.9520.9641.0000.943
    Table 4. RSROCC of different methods on LIVE 3D Phase I database
    MethodJP2KJPEGWNBLURFFAll
    PSNR0.7850.2190.9350.9160.7030.834
    SSIM0.8650.4870.9390.9190.7210.872
    Method in Ref. [1]0.9400.6410.9250.9490.7470.902
    Method in Ref. [10]0.9230.6560.9410.9510.8400.924
    Method in Ref. [9]0.9050.7290.9040.6170.5030.626
    Method in Ref. [8]0.9070.6950.9170.9680.7350.895
    Method in Ref. [15]0.8980.6320.9230.9280.8450.926
    Method in Ref. [14]0.8480.6260.9250.8990.7070.887
    Method in Ref. [16]0.9380.8060.9190.8810.7580.917
    Proposed0.9840.8700.9680.9850.9910.968
    Table 5. RPLCC of different methods on LIVE 3D Phase I database
    MethodJP2KJPEGWNBLURFFAll
    PSNR0.5970.4910.9190.6900.7300.665
    SSIM0.7040.6780.9220.8280.8340.792
    Method in Ref. [1]0.7510.8670.9230.4550.7730.728
    Method in Ref. [10]0.8470.7190.8450.8000.8500.745
    Method in Ref. [9]0.7240.6490.7140.6820.5590.543
    Method in Ref. [8]0.8670.8670.9500.9000.9330.880
    Method in Ref. [14]0.5530.5930.5930.8690.8280.825
    Method in Ref. [16]0.8640.8390.9320.8460.8600.888
    Proposed0.9291.0000.9640.9640.9290.934
    Table 6. RSROCC of different methods on LIVE 3D Phase II database
    MethodJP2KJPEGWNBLURFFAll
    PSNR0.5970.4910.9190.6900.7300.665
    SSIM0.7040.6780.9220.8380.8340.792
    Method in Ref. [1]0.7840.8530.9260.5350.8070.748
    Method in Ref. [10]0.8370.7500.8490.8270.8800.758
    Method in Ref. [9]0.7660.7860.7220.7950.6740.568
    Method in Ref. [8]0.8990.9010.9470.9400.9320.880
    Method in Ref. [14]0.6340.6470.9040.9670.8510.818
    Method in Ref. [16]0.8670.8290.9200.8780.8360.845
    Proposed0.9390.9320.9710.9870.9600.943
    Table 7. RPLCC of different methods on LIVE 3D Phase II database
    Metric3DSwIMYou(l)You(g)StSdBenoitGorleyST-SIAQNIQSVProposed
    RPLCC0.64970.75040.36500.69950.74250.70990.71330.67830.7822
    RSROCC0.56830.75670.66090.70080.75180.71960.70340.62080.8033
    Table 8. RPLCC and RSROCC of different methods on MCL-3D database
    Shuyu Huang, Qingbing Sang. No-Reference Stereo Image Quality Assessment Based on Image Fusion[J]. Laser & Optoelectronics Progress, 2019, 56(7): 071004
    Download Citation